The global transition towards electrified transportation, fundamentally driven by the need for sustainable energy solutions, has placed the EV battery pack at the core of automotive engineering. Among various chemistries, lithium iron phosphate (LFP) cells are favored for commercial electric vehicles due to their intrinsic safety and long cycle life. However, the risk of thermal runaway, a self-sustaining and uncontrollable increase in temperature, remains a critical safety challenge. Once initiated in a single cell, thermal runaway can propagate catastrophically within a densely packed EV battery pack, releasing flammable gases, intense heat, and potentially leading to fire or explosion. Traditional post-ignition fire suppression methods often struggle as they cannot reach the core of the failure inside the sealed pack. This study explores and validates a proactive, inherently safe, and cost-effective strategy: maintaining a continuous micro-positive pressure nitrogen environment within the EV battery pack to fundamentally suppress thermal runaway hazards.
The principle is rooted in the classic “fire triangle.” For a fire or explosion to occur, three elements are required: fuel (flammable gas/electrolyte), an oxidizer (oxygen), and an ignition source (an arc, hot surface, or the thermal runaway reaction itself). Within an EV battery pack undergoing abuse, the internal generation of fuel is often inevitable. Ignition sources can arise from electrical faults like arcs. The novel approach proposed here is the continuous and active removal of the oxidizer. By filling the pack’s void volume with inert nitrogen at a pressure slightly above ambient (e.g., 0.1-0.2 kPa), a passive, protective atmosphere is created. This micro-positive pressure not only displaces oxygen but also prevents moist ambient air from entering the pack during normal “breathing” cycles caused by temperature and pressure fluctuations. This addresses two major risks simultaneously: it starves any potential internal fire of its oxidizer and prevents condensation-induced insulation failure, a common secondary fault that can initiate thermal runaway. The core hypothesis is that by creating an oxygen-deficient, inert environment inside the EV battery pack, the severity and likelihood of thermal runaway and its associated fires can be drastically reduced.
Experimental Methodology for EV Battery Pack Safety Validation
The experimental design focused on comparative analysis under two distinct environmental conditions within a representative commercial EV battery pack enclosure. The test subject was a pack designed for commercial vehicles, housing 36 large-format 302 Ah LFP cells. For practicality and safety, critical experiments were conducted on a representative 1×3 cell module (at 100% State of Charge) placed inside the sealed pack. Two primary test campaigns were executed:
- Thermal Runaway Propagation Test: A heating plate was inserted between two cells to induce thermal runaway at a controlled rate ≥6 °C/min. A continuous arc igniter was positioned near the cells to simulate an internal ignition source. Two conditions were tested: (a) Pack in normal air (no suppression), and (b) Pack with continuous micro-positive pressure nitrogen purging.
- Condensation and Insulation Resistance Test: The complete EV battery pack was subjected to a harsh environmental cycle (45°C, 95% RH) with active liquid cooling. Its internal humidity and post-test insulation resistance were compared under normal conditions versus with continuous micro-positive pressure nitrogen.
Key parameters monitored included cell surface temperatures at multiple points (Ti-u, Ti-s, Ti-f, Ti-b), cell voltages, pack internal pressure and humidity, visual events (venting, fire), and insulation resistance. The criteria for declaring thermal runaway followed the standard where a cell’s surface temperature exceeds its maximum operating temperature with a rise rate ≥1 °C/s for over 3 seconds.
Quantitative Results: A Tale of Two Environments
The contrast between the uncontrolled scenario and the nitrogen-protected EV battery pack was stark and definitive, measurable across multiple physical and temporal dimensions.
Thermal Runaway Behavior and Severity
In the baseline test (normal air), the heating triggered a rapid and catastrophic failure sequence. The first cell vented at t=1008s, with ejected gases immediately igniting on the arc, leading to a violent jet fire. This event precipitated the thermal runaway of the adjacent cells within a mere 140 seconds. The EV battery pack was engulfed in flames for over 400 seconds. The recorded temperatures were extreme, with surface temperatures on the cells soaring past 600°C, indicating complete energetic material consumption.
Under the micro-positive pressure nitrogen environment, the outcome was fundamentally altered. The first safety valve opened significantly later, at t=1534s, a delay of 526 seconds compared to the baseline. Critically, the vented gases did not ignite. Only a small amount of white smoke was observed, which dissipated without combustion. The heating was stopped after venting, and the cells cooled passively. No thermal runaway propagation occurred. The maximum recorded cell surface temperature was only 148°C, well below the decomposition thresholds of LFP materials. The quantitative comparison is summarized in Table 1.
| Parameter | Normal Air (No Suppression) | Micro-Positive Pressure N₂ |
|---|---|---|
| First Venting Time | 1008 s | 1534 s (+526 s) |
| Maximum Cell Surface Temp. | 627.6 °C | 148.0 °C |
| Thermal Runaway Occurred? | Yes (3/3 cells) | No (0/3 cells) |
| Fire Duration | 413 s | 0 s |
| Pack Internal Post-Test State | Severely burned, charred module | Module intact, minor electrolyte leakage |
The governing criterion for thermal runaway is the temperature rise rate. In the normal air case, the rate skyrocketed upon venting. In the nitrogen case, the rate remained low. This can be conceptually framed as:
$$ \text{Thermal Runaway Criterion: } \left( \frac{dT}{dt} \right)_{\text{max}} \geq 1 \quad ^\circ\text{C/s} \quad \text{for} \quad \Delta t > 3 \text{s} $$
Where the experimental results show:
$$ \left( \frac{dT}{dt} \right)_{\text{max, Air}} \approx 24.6 \quad ^\circ\text{C/s} \quad \gg 1 \quad ^\circ\text{C/s} $$
$$ \left( \frac{dT}{dt} \right)_{\text{max, N₂}} \ll 1 \quad ^\circ\text{C/s} $$
Mitigation of Secondary Hazards: Humidity and Insulation
The condensation tests revealed another critical advantage of the micro-positive pressure system for a real-world EV battery pack. In a standard pack, the “breathing” effect draws in warm, humid air during cooling cycles. When this air contacts cold surfaces like cooling plates, condensation forms. This water can degrade insulation, leading to potential short circuits—a primary initiator of thermal runaway.
With continuous nitrogen purging, the internal atmosphere is controlled. As shown in Table 2 and the conceptual humidity trend in Figure 1, the nitrogen environment maintains a low, stable relative humidity (RH), effectively preventing condensation. This was confirmed by visual inspection and a vastly superior insulation resistance measurement post-test.
| Parameter | Normal Air (No Suppression) | Micro-Positive Pressure N₂ |
|---|---|---|
| Avg. Internal RH During Cycle | >60% RH (peaking ~80% RH) | ~30% RH (stable) |
| Condensation on Cold Surfaces | Yes (visible water droplets) | No |
| Post-Test Insulation Resistance | 43.7 MΩ | 847 MΩ |
| Insulation Degradation | ~95% reduction from initial state | Negligible degradation |
The insulation resistance (\(R_{ins}\)) is a direct indicator of pack health. The dramatic difference highlights the protective effect. The risk of condensation-induced failure (\(P_{cond}\)) can be modeled as being proportional to the time-integrated exposure to high humidity above the dew point inside the EV battery pack:
$$ P_{cond} \propto \int_{t} \left( \text{RH}_{int}(t) – \text{RH}_{dew}(T_{cold}(t)) \right) \, dt $$
where \(\text{RH}_{dew}\) is the dew point relative humidity at the coldest surface temperature \(T_{cold}\). The micro-positive pressure \(N_2\) system forces \(\text{RH}_{int}(t)\) to remain low, making the integrand negative or near zero, thus driving \(P_{cond} \to 0\).
Comprehensive Risk Analysis for the EV Battery Pack
The overall safety improvement for an EV battery pack employing this technology can be synthesized through a multi-parameter risk analysis framework. Risk (\(R\)) is generally defined as a function of the probability or likelihood of a hazardous event (\(L\)) and the severity of its consequences (\(S\)): \( R = f(L, S) \). The experimental data allows for a qualitative but powerful comparison across key metrics that influence both \(L\) and \(S\).
The data from three repeated trials under each condition showed high consistency, with parameter errors below 5%, confirming the reliability of the findings. A consolidated risk radar chart (described quantitatively below) visually demonstrates the superiority of the nitrogen environment. Every measured metric related to hazard severity (temperature, fire intensity) and likelihood (trigger time, humidity, insulation loss) shows improvement.
We can define a normalized hazard score (\(H\)) for the EV battery pack under a given condition based on critical parameters. A lower score indicates lower risk.
$$
H = w_1 \left(\frac{T_{max}}{T_{ref}}\right) + w_2 \left(\frac{\dot{T}_{max}}{\dot{T}_{ref}}\right) + w_3 \left(\frac{t_{vent,ref}}{t_{vent}}\right)^{-1} + w_4 \left(\frac{\text{RH}_{avg}}{\text{RH}_{ref}}\right) + w_5 \left(\frac{R_{ins,ref}}{R_{ins}}\right)
$$
Where \(T_{ref}\), \(\dot{T}_{ref}\), \(t_{vent,ref}\), \(\text{RH}_{ref}\), and \(R_{ins,ref}\) are reference values (e.g., from the normal air test or safety thresholds), and \(w_n\) are weighting factors. Using the data from our tests:
- For Normal Air: \(T_{max}/T_{ref} \approx 1\), \(\dot{T}_{max}/\dot{T}_{ref} \approx 1\), \((t_{vent,ref}/t_{vent})^{-1} \approx 1\), \(\text{RH}_{avg}/\text{RH}_{ref} > 1\), \(R_{ins,ref}/R_{ins} \approx 1\). Thus, \(H_{Air} > 1\).
- For Micro-Positive N₂: \(T_{max}/T_{ref} \ll 1\), \(\dot{T}_{max}/\dot{T}_{ref} \ll 1\), \((t_{vent,ref}/t_{vent})^{-1} < 1\), \(\text{RH}_{avg}/\text{RH}_{ref} \ll 1\), \(R_{ins,ref}/R_{ins} \ll 1\). Thus, \(H_{N_2} \ll H_{Air}\).
This mathematical representation confirms that the integrated hazard score for the EV battery pack is drastically reduced under the nitrogen atmosphere. The key metrics are compared in Table 3.
| Risk Factor Category | Specific Metric | Normal Air | Micro-Positive N₂ | Impact on Risk (L/S) |
|---|---|---|---|---|
| Severity (S) | Max. Cell Temp. | 627.6 °C | 148.0 °C | Dramatically Reduces S |
| Fire Presence | Yes (413s) | No | Eliminates Major S | |
| Flame Temp. | High (~800°C+) | Not Applicable | Eliminates S | |
| Likelihood (L) | Time to First Vent | Baseline (1008s) | +526s Delay | Reduces L |
| Internal Humidity | High (>>60% RH) | Low (~30% RH) | Reduces L of Short Circuit | |
| Insulation State | Degraded (43.7 MΩ) | Robust (847 MΩ) | Greatly Reduces L | |
| Propagation | Full (3/3 cells) | Arrested (0/3 cells) | Reduces L of Cascading Failure |
Discussion and Engineering Implications
The mechanism of suppression is twofold. First, by reducing the oxygen concentration within the EV battery pack to very low levels, the inert atmosphere prevents the combustion of vented flammable gases (e.g., \(H_2\), CO, hydrocarbons) and the burning of ejected electrolyte. This directly tackles the most severe consequence—fire. Second, the absence of combustion removes a powerful feedback mechanism. In a normal air event, the external fire acts as a massive external heat source, rapidly heating neighboring cells and guaranteeing propagation. In the nitrogen-inerted EV battery pack, the venting event is primarily an endothermic or mildly exothermic gas-release process without a superimposed flame, allowing the system to dissipate heat and avoid triggering neighbors. Furthermore, the micro-positive pressure acts as a one-way barrier against ambient moisture, solving a chronic reliability issue for packs operating in humid climates and indirectly preventing one common root cause of internal short circuits.
From an engineering perspective, this system is attractive for several reasons. Nitrogen is abundant, inexpensive, and inert. Maintaining a micro-positive pressure (0.1-0.2 kPa) requires minimal energy and simple components: a small compressor or stored \(N_2\) cartridge, a pressure sensor, and a control valve. This system is always active, requiring no “detection” of a fault to initiate, thus acting as a preventative barrier rather than a reactive countermeasure. It is particularly well-suited for LFP-based EV battery packs because, unlike some nickel-rich chemistries, LFP cathodes do not release oxygen during decomposition, meaning the primary oxidizer inside the pack is the ambient air we exclude.
In conclusion, this experimental study provides compelling evidence that maintaining a continuous micro-positive pressure nitrogen environment within an LFP EV battery pack is a highly effective, practical, and dual-purpose safety technology. It actively suppresses thermal runaway severity by preventing internal fires and reduces the likelihood of failure initiation by controlling internal humidity and preserving insulation integrity. This proactive approach to managing the internal atmosphere of the EV battery pack represents a paradigm shift from fighting fires to preventing the conditions that allow them to occur, offering a robust and integrable solution for enhancing the safety of commercial and high-capacity energy storage systems.